Data-Driven Approach for Passenger Mobility Pattern Recognition Using Spatiotemporal Embedding

Table 2

Characteristics of mobility patterns.

ID	Proportion (%)	Spatial features		Temporal features			Possible activity
		The origin station	The destination station	The start time	The day of week	Travel time
C1	13.716	Mainly residential POIs	Mainly entertainment, working, hospital, and shopping POIs	Mainly 7–8	Mainly weekdays	Mainly 40 min–80 min	Working (long distance)
C2	19.817	Mainly entertainment, working, shopping, and education POIs	Mainly residential POIs	After 17	Weekdays and Sundays	Within 40 min	Home (short distance)
C3	13.908	Mainly residential, entertainment POIs	Mainly entertainment and shopping POIs	9–19	Mainly weekends	Within 60 min	Entertainment and shopping
C4	14.506	Mainly entertainment, working, and shopping POIs	Mainly residential POIs	Mainly 17–19	Weekdays and Sundays	Mainly 40 min– 80 min	Home (long distance)
C5	22.092	Mainly residential POIs	Mainly entertainment, working, shopping, and POIs	Mainly 7–9	Mainly weekdays	Mainly within 40 min	Working (short distance)
C6	15.961	Mainly entertainment, shopping, and hospital POIs	Mainly entertainment, shopping, and residential POIs	Mainly 11–17	Weekdays	Mainly within 40 min	Others